NVIDIA CUDA GL Reviews & Product Details

- Containerizing a graphical program with ML backend using CUDA has become super easy, thanks to NVIDIA's effort of glvnd. With Cuda GL containers, it is much easier to get up and running within a few minutes, avoiding all the headache of incompatible libraries versions and arbitrary crashes. Nvidia toolkit has support for a large range of ml/dl libraries Review collected by and hosted on G2.com.

- From the last time I saw it, I remember the development and release of new containers were paused for some other tech debt work. Other than that, I think community around is good and slowly growing. Review collected by and hosted on G2.com.

I used CUDA in college to study stuff like parallel processing with GPUs so it's good if you want to know the basics and then start off with more advanced steps Review collected by and hosted on G2.com.

The only disadvantage seems to be its age, it's good for someone to clear the basics but it's kind of outdated and is load intensive . i had to run its program on google collab rather than local pc due to too much time taking to run Review collected by and hosted on G2.com.

-The best part about it is the cross platform support which ensures that it behaves in same manner throughout different operating systems.

-It is quite easy to integrate it with already built model and is quite helpful in speeding up compute heavy models/applications . Review collected by and hosted on G2.com.

This is not necessary a con but building basic implementations might be easier but for more complex applications you require good level of understanding , but then again that part is usually taken care by senior devs.

So it might take some time to get a good command over it upto and intermediate to expert level. Review collected by and hosted on G2.com.

I worked on NVIDIA CUDA long time back in 2010. I was building a coding assessment platform similar to Leetcode but for NVIDIA CUDA, which tracks the number of cores and memory usage along with correctness of the solution. I found it difficult to understand, as the concepts are different. However I was able to get the project working and it was used by the university of California to access the skill of the students in parallel programming.

What I liked best was the ability to split and parallelly execute programs which allows massive throughput and reduced time to run compute-intensive programs by orders of magnitude. Review collected by and hosted on G2.com.

As it was new, I had difficulty understanding the concepts of designing programs to run on multiple cores. I also did not like the tooling around it. There is no debugging or good IDE support. This was initially when I worked on it way back in 2010. Not sure of the current state. Review collected by and hosted on G2.com.

I have been using NVDIA from 6 years now and it has been very good in performance. Durability at its best.Running smoother for 6 years is something which should be appreciated. Review collected by and hosted on G2.com.

There is nothing I dislike regarding NVIDIA, but given technology advancements, products should be available for reasonable prices. Apart from that I think everything is good. Review collected by and hosted on G2.com.

It has the best in market performance when in comes to computing. They provide easy to use architecture and for distribution purposes. Nvidia toolkit is very comprehensive. Review collected by and hosted on G2.com.

There is no such specific downside but everything has a scope of improvement. It is designed for mostly older cards though suits well for newer ones too. Overall great. Review collected by and hosted on G2.com.

I like several aspects of NIVIDA CUDA GL:

1. It speeds up the tensor computation significantly.

2. It has better compatibility with a different version of GPU.

3. It has reduced OS (i have experience on Ubuntu 18.04 and Ubuntu 20.04) crash and bug in some library after installing the new library. Review collected by and hosted on G2.com.

.Till now I have not found any issue while working with NIVIDA CUDA GL. Everything is great. Review collected by and hosted on G2.com.

It's speed and parallel computing features. Review collected by and hosted on G2.com.

It's cost is a bit high for average users like me. Review collected by and hosted on G2.com.

Ease of use and easy adoption across NVIDIA GPUs. The CUDA language also seems familiar, so it's easy and quick to understand and learn as well which is one of the biggest advantages it has over it's competitors. Review collected by and hosted on G2.com.

The only problem is with it being specific to NVIDIA's architecture, it's a bit difficult to recommend considering there are alternatives like OpenCl. If there's any improvements to be made, I think try to be more inclusive? Review collected by and hosted on G2.com.

We are already aware about programming languages, CUDA is as simple as any other programming language to understand and quickly code the parallel processing logic. Review collected by and hosted on G2.com.

Lack of online resources and quick tutorials with examples must be available to grow the usage. Also, understanding when to use and when not to use CUDA is a skill, that also should be explained. Review collected by and hosted on G2.com.